Effective and efficient pixel-level detection for diverse video copy-move forgery types

Copy Move Forgery Detection: The Current Implications and Contemporary Practices

“…The experiments show that, when validated against the current state-of-the-art methods or the aforementioned conventional approaches, the proposed method in the original paper [31] achieves the most elevated detection accuracy, the lowest false positive rate (FP), and improves at least '16 percent' and '8 percent' of F1 scores on the GRIP 2.0 dataset and a mixture of 'SULFA 2.0' and 'REWIND' datasets, respectively. The suggested technique, in addition, has a low computational time (4.45 s/Mpixels), which is approximately half to three-quarters the time required by the most recent 'DFMI-BM (8.02 s/Mpixels) and PM-2D (13.1 s/Mpixels) approaches' [31]).…”

“…Compared to standard approaches the paper [31] implemented the related and corresponding techniques on a stream of images for video CMFD. A critical and very tough problem, video copymove forgery detection (VCMFD) is made more complex by a number of challenges, including a large quantity of video information, a range of forgery kinds, and a rich set of forgery objects, and homogeneous forgery sources.…”

“…A critical and very tough problem, video copymove forgery detection (VCMFD) is made more complex by a number of challenges, including a large quantity of video information, a range of forgery kinds, and a rich set of forgery objects, and homogeneous forgery sources. These obstacles pose four major issues that have not yet been handled in VCMFD: Forgery videos are frequently detected with the ineffectiveness in several popular forgery cases, as well as with inefficient matching in computing a multitude of video pixels with multi-dimensional elements under dozens of corresponding iterations, as well as with high false positive (FP) rates [31].…”

“…This technique depends on the inherent qualities of precise keypoint matches. ; iii) responsive block stuffing based on absolute keypointmatches contributes to the valid and a computationally efficient filling of suspicious regions, even at high frame rates; iv) a novel coarse-to-fine filtering approach based on the inherent properties of precise keypoint matches is proposed to more efficiently calculate the needed processes [31]. After that, the suspicious area locations that are associated with the forgery vision permanence concept suggest the presence of forgery videos [31].…”

A comprehensive taxonomy on multimedia video forgery detection techniques: challenges and novel trends

“… [ 97 ] 2022 The proposed approach used the Farneback method. Up to 98% The proposed method gives better accuracy, but computational complexity is one of the big challenges of this scheme [ 233 ] 2022 A new proposed scheme for copy-move detection at pixel-level by using fast keypoint-label matching NA However, the proposed method is still a kind of craft method that depends on many hyper parameters that affect final performance. …”